sumetho - Fotolia
Cloudian Inc. added intelligence to its HyperStore software this week by fully integrating its object storage with Apache Hadoop analytics.
The Hadoop support in HyperStore 5.1 gives customers a key feature that is needed to turn big data into smart data.
The integration allows customers to use HyperStore software and appliances to run in-place analytics for Hadoop without requiring Hadoop Distributed File System (HDFS).They can analyze data right on the storage.
This data analytics capability will be initially offered for Hortonworks Hadoop environments.
"We added smart data analytics," said Paul Turner, Cloudian’s chief marketing officer. "Because we are an object-based storage platform, we have both the data and the metadata, so we are able to run Hadoop analytics against the data that Cloudian collects."
The typical way to mine data is to extract and load it onto HDFS. Hadoop makes three copies of data for protection, which can result in a lot of extra storage for large implementations. Integrating HyperStore with Hortonworks Hadoop can eliminate that redundancy.
"Most people are copying their data set to HDFS and then running these queries," said Jim Bagley, senior analyst at Storage Strategies Now. "With Cloudian’s capability, they can run it directly."
Cloudian’s Amazon S3 compliance allows HyperStore customers to use the Amazon Web Service (AWS) cloud as a tier for long-term data behind a Cloudian implementation.
"I don’t know of anyone who has implemented [Hadoop] the way Cloudian does," Bagley said. "They are using the same interface points of S3. I think they had to do quite a bit of engineering overall. It's not a small task. [Hadoop] is a byproduct of this. They replicate the functions of S3 with their own software."
Cloudian also added federated geo-replication so HyperStore can distribute data over long distances efficiently. The startup also introduced its Smart Support Analytics program, which includes a call home feature and the ability to support end users' credentials in the cloud for better security. Cloudian is supporting OpenStack Icehouse with the 5.1 release.
"Enterprises expect you to follow their security credentials," Cloudian's Turner said.
The management benefits of using an object storage system
Hadoop clusters aren’t ideal for all environments